How many percentage points more than pot odds you need, and who pays the most.
In a cash game a call is good when your equity beats the pot odds. In a tournament with payouts, the chips you lose cost more prize money than the chips you win are worth, so you need more. The risk premium is that extra, measured in percentage points of equity.
Take your ICM equity (your share of the prize pool) in three situations: if you fold (F), if you call and win (W) and if you call and lose (L). The equity you need for the call to be as good as folding is (F − L) / (W − L). The pot-odds number in chips is what you would need with no payouts. Their difference is the risk premium.
It is close to, but not the same as, the bubble factor. The bubble factor is a ratio, (F − L) / (W − F). The risk premium is a difference in equity, and it is the number you compare directly with the equity of your hand. Sources often use the two names loosely, so check which one a number refers to.
Nine players are left, the payouts are 28/19/13/10/8/7/6/5/4 percent, the blinds are 50/100 with a 12.5 ante, so 262.5 chips are in the middle. The stacks behind are 3000, 2500, 2000, 1600, 1300, 1000, 800, 500 and 300. The 3000 stack goes all-in and you, with 1600, are next to act and not in the blinds.
A hand that has 55% against his range is an easy call in chips and a fold under ICM. Open these numbers in the ICM calculator to check every figure above.
The table shows the risk premium, in points, for the same final table when you are the row stack and call the shove of the column stack. Nothing else changes.
| You \ shover | 3000 | 2500 | 1600 | 1000 | 800 |
|---|---|---|---|---|---|
| 3000 | – | 13.6 | 7.1 | 4.2 | 3.3 |
| 2000 | 18.5 | 17.9 | 11.4 | 6.2 | 4.8 |
| 1300 | 17.0 | 16.6 | 15.5 | 9.7 | 7.3 |
| 800 | 15.1 | 14.7 | 13.9 | 12.9 | – |
| 500 | 12.7 | 12.5 | 11.8 | 10.9 | 10.4 |
The pattern is easy to read. The premium is largest for a medium stack facing a bigger one (for example 2000 against 3000: 18.5 points) and smallest for a chip leader calling a shorter stack (3000 against 800: 3.3). The chip leader risks little prize equity, while the medium stack risks a lot to gain a little. A short stack of 500 pays a smaller premium than a medium stack because it already holds little equity to lose.
Same spot (you, 1600, calling the 3000 shove), four payout structures:
| Payouts | Equity needed in chips | Equity needed in ICM | Risk premium (points) |
|---|---|---|---|
| Winner takes all | 46.2% | 46.2% | 0.0 |
| Top-heavy (38/24/15/9/6/4/2/1/1) | 46.2% | 60.8% | 14.6 |
| Typical final table (28/19/13/10/8/7/6/5/4) | 46.2% | 63.9% | 17.7 |
| Flat (20/16/13/11/10/9/8/7/6) | 46.2% | 69.2% | 23.0 |
With winner-takes-all the premium is exactly zero: equity is proportional to chips and ICM agrees with chip EV. The flatter the ladder, the more it pays to survive and the larger the premium. Do not generalise that into "steeper payouts always mean tighter play": the effect depends on where you are on the ladder and on who covers whom.
The ICM trainer shows the risk premium for every spot you play, together with the bubble factor and the value of your stack after a fold, a win and a loss.
One short lesson a day, each with a worked example in exact numbers: chips vs. money, the bubble factor, who feels the pressure, payout structures, satellites. No account needed.
There is no good or bad value. It is the price of risking your stack. In our final table example it ranges from about 3 points for a chip leader to about 18 points for a medium stack facing a bigger one.
No. Bubble factor is a ratio of the equity you lose to the equity you gain. The risk premium is the difference between the equity ICM requires and the pot-odds equity. They move together but are not the same number.
It is exact for the Malmuth-Harville ICM model, which is an approximation of real tournament results. Treat it as a strong guide rather than a fixed rule.
progressGTO guides use our own solution book (v17d, 8-max tournaments with a 12.5% ante from each player, chip EV): an approximation of equilibrium with a modelled postflop, not an exact solution. Numbers are computed from the book and from the exact ICM model.